Parallel and distributed programming is intrinsically more difficult then sequential programming, yet few effective tools or methodologies have been developed to help programmers understand the behavior of their parallel programs. Browsing source code and tracing program output are tedious and ineffective approaches for parallel program understanding. Program visualization, which relates a program's behavior to the programmer's model of the system's components and interactions, has been shown to be a novel and highly effective approach to program and algorithm comprehension. Extending and adapting program visualization to parallel programming can aid comprehension of the complex concurrent events and transitions that occur in parallel programs. This project will define a model for the capture and display of parallel program events and transitions, based upon the path- transition paradigm for animation, and partial ordering of events. The model will be used to develop a prototype for visualizing parallel programs. The model and its prototype will be tested upon a suite of scientific parallel programs.
